Anda di halaman 1dari 5

ME 457 Experimental Solid Mechanics (Lab)

Torsion Test : Solid and Hollow Shafts

The purpose of torsion testing usually parallels that of uniaxial tension tests. From the
experiment, the shear elastic modulus (G), shear proportional stress (τp), shear yield
stress (τy), and the stress-strain behavior in general, can be obtained. However, in contrast
to uniaxial tension tests, the stresses are not distributed uniformly over the cross section.

Each test will be conducted until failure i.e., it will end in the buckling of the hollow
specimen or fracture for the solid specimen. Thus, this experiment will also allow
observing the different modes of failure for solid and hollow circular shafts made of
ductile or brittle materials.

Torsion loading results in twisting of one section of a body with respect to a contiguous
section. During the test the angle of twist φ and the applied torque T are measured as the
test proceeds. For a circular cross-section, in the absence of the other loads, pure shear
stress state exists at each point. Torsional elastic shear stresses vary linearly from zero at
the axis of twist to a maximum at the extreme fibers. Thus, in a solid circular bar, when
the surface fibers reach the yield shear stress they are, in a sense, supported by elastic
interior fibers. Consequently, the elastic resistance of the remainder of the section masks
the effect of yielding of the surface fibers during their early stage of yielding. Usually, it
is not until considerable yielding has taken place that any noticeable effect of nonlinearity
is apparent using a simple mechanical troptometer to measure the angle of twist φ
(calibrated in increments of 0.2 degrees). Therefore, it is practically impossible to
determine when the extreme fibers of the solid specimen in torsion start to yield. This
difficulty is overcome by the use of hollow (thin-walled) specimens, which give more
sensitive measures of the elastic-plastic transition since all the fibers are at about the
same stress. However, for thin-walled tubes with large ratios of diameter vs. thickness
(D/t > 10) there is a tendency for premature local buckling failure to occur. Therefore the
actual dimensions of the specimen used must be carefully chosen.

To test the material in torsion the proper test procedure is needed. It involves mounting a
shaft into the testing machine, applying torque incrementally and measuring both the
applied torque and the corresponding angle of twist. Using the appropriate formulae,
relationships and the measured dimensions, we can determine the shear stress and shear
strain on the shaft. Then, one can plot the torque vs. angle of twist, and shear stress vs.
shear strain from which one can find the material properties previously mentioned.

The assumptions made in this experiment include but are not limited to the following:
1. The torque is applied along the center of axis of the shaft.
2. The material is tested at steady state (absence of strain rate effects).
3. Plane sections remain plane after twisting (the circular section conforms to this

1. Torsion testing specimens:
• Solid (ductile) aluminum shaft,
• Hollow (ductile) aluminum shaft (both ductile specimens are made of
the same material),
• Solid (brittle) aluminum shaft.
2. Caliper
3. Troptometer
4. Torsion testing machine
5. Measuring tape
6. Safety glasses

Test Procedure (for solid and hollow ductile shafts):

1. Measure the diameter of the test specimen using the caliper (take an average of 5
2. Slide the shaft into the troptometer and tighten down the two fastening screws to the
3. If the hollow shaft of aluminum is used, slide two steel inserts into the ends where the
sample will be clamped down.
4. Clean the clamps used to hold the shaft in place.
5. Insert the shaft into the right clamps’ chucks in the testing machine. Tighten the
clamp ensuring that the grip is very tight.
6. Slide the other clamp over the other end of the shaft, and tighten down tightly as well.
7. Adjust the troptometer to the correct position from both ends (i.e. do not let it contact
with any part of the testing machine.).
8. Measure the length between the troptometer clamps Lt using the tape measure. This
length will be used in calculations later.
9. Next, measure the length between the torsion machine's clamps, Lc.

Note : While inserting the specimen into the clamps, move the weighing house to the right
until close to or stopped by the shaft end bearing against the inside of the chuck.

10. Make zero the torque indicator by aligning the red needle with the black one on the
torque indicator. Also, use the zeroing lever to make sure that the needles point to
11. Make zero the troptometer indicator by hand.
12. Turn the crank manually forwards (counter-clockwise) and observe the troptometer
indicator. You want to increase the angle of twist by 1° increments up to12°.
13. At each interval as describe in the above step, read the measurement (using the red
needle) displayed by the torque gage. NOTE: The torque gage has to be read from 0
to 10,000 in-lb. Hence, every reading that you get on the regular scale must be
multiplied by 10 for the correct results.
14. Also take the reading off the scale on the right (rotating) clamp housing. This should
be done at every even degree that is reached by the troptometer.(i.e. at 2°, 4°, 6° etc.)
15. Construct the following table and tabulate the torque T, and the corresponding angle
of twist of the troptometer φt, and clamp φc:

Torque (T) Angle of twist of Angle of Shear stress Shear strain
[in-lb] troptometer, (φt) twist of (τ) [psi] (γ) [in/in]
[°] / [rad] clamp, (φc)
[°] / [rad]

16. Once 12° is reached, unload the specimen by turning the crank backwards
(clockwise) and take the torque gage reading and the angle of twist of the clamp at 1°
intervals (read off the black needle this time). All the while, watch the torque gage
making sure that it has not reached zero yet. Tabulate this data as a continuation on
the same table used in step 15.
17. Watch for the torque gage to read zero. At this point, take down the troptometer
reading and the angle of twist of the clamp.
18. Next reload the specimen by turning the crank manually forwards (counter-
clockwise) and observe the troptometer indicator. You want to increase the angle of
twist by 1° increments up to the final angle of 20°.
19. At each interval as describe in the above step, read the measurement (using the red
needle) displayed by the torque gage. NOTE: The torque gage has to be read from 0
to 10,000 in-lb. Hence, every reading that you get on the regular scale must be
multiplied by 10 for the correct results
20. When φt = 20o totally loosen the troptometer and let it hang loosely and twist the
specimen to failure. The hollow aluminum shaft is considered to have failed when it
Note the angle of twist that was needed to fracture or buckle the specimen from the
angle of twist indicator located on the right rotating arm of the torsion machine, (φc)f.
It is best to count the number of rotations and then later convert it to degrees when
writing the report. Also, record the final torque reading of the specimen at failure, Tf.

Test Procedure (for a brittle shaft):

1. Insert the brittle (cast aluminum) shaft into the machine's clamps and tighten them.
Twist the specimen until failure and observe the failure mode.

1. For two ductile shafts (hollow and solid) complete the above table by using the
following equations. Use the torque and the troptometer angles of twist φt (not the
angle of twist of the clamp φc, it will be used later to calculate the fracture strain γf =
(φc)f r/Lc ):

(a) τ = T r/Ip - valid only in the elastic range, for both solid and hollow shafts.

(b) τavg = T/2π(ravg)2t - valid only for the hollow shaft in the elastic and inelastic ranges.

(c) γ = φt r/Lt - valid in the elastic and inelastic ranges for both shafts.

In order to find the actual shear stress for the inelastic loading portion of the solid
shaft (since there are no equations to calculate them), continue on with step 2 and plot,
first the shear stress vs. shear strain diagram for the hollow shaft. Using this plot, cross-
referenced the shear stress from the calculated shear strains attained from the solid shaft
data (see hint on page 5). We are able to do this because the stresses are nearly uniform
throughout the wall thickness of the hollow shaft as compared to the solid one (both
shafts are made of the same material). The cross-referenced shear stress can then be
tabulated in your table as the “actual” shear stresses at the surface of the solid shaft.

2. Plot the graphs:

• Torque (ordinate) vs. angle of twist (abscissa)
• Shear stress (ordinate) vs. shear strain (abscissa)
There should be 4 graphs in all, two for each specimen.

3. From the graphs determine:

• Proportional shear stress (τp) for the hollow shaft.
• Yield shear stress based on a 0.2% offset (τy) for the hollow shaft.
• Shear modulus of elasticity (G) during loading and unloading for both shafts.
• Residual shear strain (γr) for both shafts
• Next, starting at the point of residual strain γr, calculate the new proportional
limit (τp), yield point (τy) based on a 0.2% offset, shear modulus of elasticity
(G). In addition, calculate the fracture strain (γf) and corresponding fracture
strength (τf) in torsion. Compare and discuss in detail the differences between
these new: τp, τy and G values and those obtained in the beginning of the
torsion test experiment. Explain any observed differences concisely.

• The point of inflection at the end of the elastic portion of the loading curve is where
τp is located.
• Where the loading curve meets the 0.2% offset line is where you will find τy.
• The shear elastic modulus, G, usually is not obtained as the shear stress, τ, divided by
the corresponding shear strain, γ, but rather the slope of the linear portion of the
stress-strain (τ-γ) curve. This slope is calculated by dividing the change of shear
stress, Δτ, by the corresponding change of shear strain, Δγ, in each linear range of
loading and unloading case from the stress-strain curves, i.e. G = Δτ/Δγ. The shear
stress vs. shear strain graph should have two curves: one corresponding to the initial
loading up to 12o with subsequent unloading to zero torque and the other reloading
curve which starts from the residual shear strain (γr) up to 20° and then to failure, both
drawn on the same plot (vary them by color or some other means).
• The residual shear strain, γr, is determined by measuring the strain between the zero
shear strain and the value after unloading to zero torque. (For optional extra credit:
the residual shear stress in the outside fibers of the solid shaft, τr, can be estimated
as: τr = τmax - Δτ, where due to elastic unloading Δτ = ΔT r/Ip = Tmax r/Ip.)

4. Discuss the failure modes for all three specimens based on the relation between pure
shear and normal stresses (see the helpful figure below).
a) Explain why, in torsion test, the ductile solid shaft sheared on a right section
whereas the brittle (cast aluminum) solid shaft fractured on a helicoidal path.
b) Explain why the brittle cast aluminum specimen fractured at a “negative-
slope” 45° angle while the ductile hollow aluminum specimen buckled at a
“positive-slope” 45° angle. (The figure below should be a good hint.)
c) Why did the hollow specimen experience local-buckling (the twisting shape at
failure) under torsion? What would happen to the twisting angles of local-
buckling if the torsion was done in the opposite direction?

5. It was observed that we had to use cross-referenced shear stresses from the hollow
shaft τ vs. γ curve to predict the “actual” stresses on the surface of solid shaft during
inelastic loading. We want to investigate the difference between this “actual” shear
stress and the one we would get from equation (a) in step 1 above. Plot the shear
stress vs. shear strain graphs for the ductile solid shaft again but this time draw two
curves: one based on the cross-referenced shear stresses and the other based on the
shear stresses obtained from equation (a) of step 1 (see hint below). Compare the two
and give a thorough analysis on why there is a difference between them.

Hint: How the cross-referenced shear stresses can be determined.


γ γ
same γ